JPH04167104A - Robot work planning device - Google Patents

Abstract

PURPOSE:To attain an effective holding attitude of a robot by deciding the holding attitude of robot based on the distance obtained by a distance conversion means between the open space drills and a holding attitude candidate obtained by a holding attitude candidate calculation means. CONSTITUTION:A holding attitude candidate calculation means 2 is provided together with a stable attitude calculation means 3, a rotational symmetry calculation means 4, a description production means 5 for space near a holding position, a description analyzing means 6 for space near a holding position, and a reholding place deciding means 7. In such a constitution, a holding attitude candidate is previously obtained and only the one coincident with this candidate is selected out of plural open space drills. Thus the selected space drill can always be held and the holding direction is effectively selected. In addition, a pick-and-place work plane can be effectively carried out.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a work planning device for a robot that plans work such as grasping and moving a target object.

[Prior Art] Conventionally, a robot automatically determines an action plan for grasping, moving, and placing (big-and-place) a target object based on the initial state of the target object and the target state after assembly. A work planning device is known.

Here, one of the work plans of the robot is a plan for grasping a target object. In this grasping plan, considering the approach to grasping posture and the ease of movement after grasping,
It is considered best to grasp the object from the direction of a wide space as far away as possible from surrounding obstacles.

Therefore, the present inventors have proposed a grasping planning method that makes it possible to grasp a target object from a wide spatial direction during both pick and place operations during pick-and-place work by using spatial descriptions near the grasp position (see Section 1). 5th Academic Conference of the Robotics Society of Japan (November 1986).

In this example, a spatial description of the vicinity of the target object in the initial state and target state of the target object is performed. This spatial description is
This is done by dividing the space near the target object into pentagonal pyramids (space pyramids) centered at the grasped position, and examining the interference between surrounding objects and all the space pyramids. In other words, the length from the apex of the space cone (grasping position) to the point where it first intersects with surrounding objects is greater than the distance at which the robot's grasping part (hand) can exist and approach operations are possible, that is, it is an open space cone. demand.

Next, this open space cone is found in both the initial state and the target state, and a space cone common to both, that is, a common open space cone is found. Then, in order to investigate the distribution of common open space cones, distance transformation is performed for all common open space cones. This distance conversion is performed by finding the shortest distance from the space cone that did not become a common open space cone for each common open space cone, and the direction of the open space cone that is larger than this distance conversion value is the space that is not occupied by the object. It shows that the area is wide or wide.

Therefore, considering the stable grasping of an object by two parallel fingers, by selecting the one with a large distance conversion value among the objects where the grasping center plane formed at the center of the two parallel surfaces of the object and the common open space cone intersect, The grasping posture is determined.

According to the above-described conventional example, it is possible to grasp the object from a direction in which a wide space exists, as far away as possible from surrounding obstacles, and it is possible to determine a work plan for pick-and-place work that will cause the least number of problems.

[Problems to be Solved by the Invention] However, in the conventional method of determining a work plan, when selecting a grasping posture for an object, only the central plane of grasping is considered in relation to the relationship between the object and the robot's hand. do not have. Therefore, after determining the grasping posture as described above, it is determined whether or not grasping is actually possible, taking into consideration the placement position and shape of the robot's hand and the target object. For this reason, there is a problem in that the method of avoiding interference between the object and the hand requires trial and error, and the determination of the plan becomes inefficient.

Further, in some work plans, there may be cases where it is not possible to directly transition from the initial state to the target state, and the target object must be placed once, changed hands, and then transitioned to the target state.

In the conventional method, in this case, a common open space cone must be created for all combinations of the states of the target object, which has the problem of inefficiency.

The present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide a robot work planning device that can efficiently select grasping postures that can be actually grasped.

[Means for Solving the Problems] A robot work planning device according to the present invention determines the position of a predetermined grasping position of an object based on the shape of the object and the shape and movement direction of the object grasping part of the robot. a grasping posture candidate calculating means for calculating grasping posture candidates that are postures that allow grasping of an object; an open space cone selection means for selecting an open space cone that does not interfere with surrounding objects, a distance conversion means for calculating the distance from the surrounding objects for each of the obtained open space cones, and a distance conversion means for determining the distance from the surrounding objects for each of the obtained open space cones, and The present invention is characterized by comprising grasping posture determining means for determining a grasping posture based on both the distance in the open space cone and the grasping posture candidates obtained by the grasping posture candidate calculation means.

[Operation] The robot work planning device according to the present invention has the above-described configuration, and obtains grasping posture candidates in advance. Then, only those cones that match the grasping posture candidates are selected from among the open space cones. Therefore, the selected spatial cone is always graspable, and the grasping direction can be efficiently selected. In addition, the pick-and-place grasping posture is determined only from the open-space cone that matches the grasping posture candidate without finding a common open-space cone from the open-space cones in the two states, making the pick-and-place work plan more efficient. It becomes possible to do so.

[Embodiments] Hereinafter, embodiments of the present invention will be described based on the drawings.

FIG. 1 is a block diagram showing the overall configuration of a work planning device according to this embodiment.

The shape of the target object, the position and orientation of the initial state and target state,
Conditions such as the state of surrounding obstacles are described in the three-dimensional model 1 of the environment. Then, using the data of the three-dimensional model 1 of this environment, various calculations are performed to create a work plan, but in this example, grasping posture candidate calculation is performed to find a posture that can be grasped from the shape of the target object. means 2, stable posture calculation means 3 for determining a state in which the target object can be placed stably on a table; rotational symmetry calculation means 4 for calculating the rotational symmetry of the target object; grasping for creating a geodesic dome in the vicinity of the grasping position; A positional vicinity spatial description creation means 5, a grasped positional vicinity spatial description analysis means 6 that performs analysis such as distance conversion from this geodesic dome, and a holding position determining means that determines a holding position in a wide space on the table when changing hands is necessary. 7.

The data obtained through these analyzes are then fed back to the three-dimensional model 1 of the environment, thereby creating data for work planning.

A work plan creating means 8 is connected to the three-dimensional environment model 1, and the work plan creating means 8 creates a work plan from an initial state to a target state.

Further, the work plan creating means 8 has a state transition determining means 9 therein, and this state transition determining means 9 determines the transition of each individual state.

Here, FIG. 2 shows an overall flowchart of the operation of the robot work planning device according to this embodiment.

In this way, first, a three-dimensional model 1 of the environment is created from data such as the initial state, target state, and surrounding environment (Sl). Next, from the shape of the target object in this three-dimensional model 1, the grasping candidate posture calculation means 2 calculates a grasping candidate posture (S2), the stable posture calculating means 3 calculates a stable posture (S3), and the rotational symmetry calculating means 4 calculates a grasping candidate posture. Perform rotational symmetry calculation (S4),
This result is described in three-dimensional model 1. .

Next, the work plan creation means 8 creates a work plan using the data of the three-dimensional model 1 (S5), but at this time, the state transition determination means 9 activates the grasped position vicinity spatial description creation means 5, Data for discrimination is described in three-dimensional model 1. Then, while determining state transitions, a solution to the work plan from the initial state to the target state is found by searching (S8)
. In addition, if it is necessary to change hands in the work plan,
The holding changing position is determined by the holding changing position determining means 7 (S6).

Below, the contents of each step will be explained in order.

Grasping Posture Candidate Calculation The operation of the grasping posture candidate calculation means 2 will be explained based on FIG. 3. For example, consider grasping posture candidates for an L-shaped object as shown in FIG. 3(A).

First, determine the gripping position, but since any number of gripping positions can be set, the distance from the surrounding edge of the object,
This is determined by setting the interval between grasped positions to a predetermined value in advance. In this example, four points (indicated by X marks in the figure) have been determined as shown in FIG. 3(A).

Next, the direction in which the robot's hand can grasp the object is represented by a straight line around this grasping position. That is, for the four grasping positions described above, FIGS. 3(B) to (E)
Grasping posture candidates will be calculated. In this example,
There are a total of 138 grasp posture candidates.

Note that the distance between the straight lines is a predetermined distance. −
In this way, grasping posture candidates can be found from the shape of the hand, the shape of the target object, and the like.

When changing the holding of an object for stable posture calculation, the object must be placed on a horizontal table.

Therefore, the stable posture calculation means 3 calculates a stable posture that can be stably placed for this change of grip.

First, consider five envelopes that enclose the target object without creating a concave portion, and consider perpendicular lines drawn from the center of the object to each surface of these five envelopes. If the foot of this perpendicular line is within the plane of the five packages, it can be placed stably with that plane in contact with the table surface.

For example, for the L-shaped object shown in FIG. 3, six stable postures as shown in FIGS. 4(A) to 4(F) are required. Then, serial numbers are assigned to these determined stable postures.

Calculation of rotational symmetry Here, if the target object has a rotationally symmetrical shape, a model that does not recognize this rotational symmetry will treat the same grasping posture as different objects.

That is, when recognizing a target object, usually X.

Recognizes using Y and Z three-dimensional coordinates. Therefore, Figure 5 (
In the three states A) to (C), if the axis indicated by the arrow is a specific axis (for example, the X axis), all three states will be recognized as different.

In this way, if the symmetry of the target object is not considered,
Depending on how the recognition system takes the object coordinate axes, even the same state may be recognized as different states. Therefore, as shown in FIGS. 6(A) to 6(C), even if the grasping postures are substantially the same, they are recognized as different. Therefore, depending on how the coordinate axes are taken at the time of big and how the coordinate axes are taken at the time of place, a grasping plan that is actually possible may be determined to be impossible.

Additionally, since substantially the same stable postures as shown in Figure 5 (A), (B), and (C) are recognized as different, the number of stable postures increases and the number of paths in the work plan decreases. This results in unnecessary searches.

In this embodiment, the rotational symmetry of the object is derived and the relationship between grasping postures is described. Since this description determines the identity between grasping postures, it is also possible to plan for changing the grip of a target object with rotational symmetry. Further, since identity is determined based on the rotational symmetry of the stable posture, unnecessary searches can be omitted.

Rotational Symmetry Calculation Algorithm A specific algorithm for describing this rotational symmetry is shown below.

(1) Transform one surface of the target object to rotate around the center of gravity of the target object, and obtain a transformation that overlaps the surface of the target object (including the surface itself).

That is, the following four points are checked as an initial check.

(a) The area is the same.

(b) Same perimeter.

(c) The distance from the center of gravity of the target object to the center of gravity of the surface is the same.

(d) The angle between the vector from the center of gravity of the object toward the center of gravity of the surface and the normal vector of the surface is the same.

And for those that satisfy (a) to (C),
Find a transformation where the surfaces overlap as follows.

If the angle formed by the vector from the center of gravity of the object toward the center of gravity of the surface and the normal vector of the surface is other than O'', the rotation transformation around the center of gravity of the object that may overlap the surface is uniquely There is only one.

Therefore, the overlapping surfaces are determined by this rotational transformation, and the symmetrical transformation is determined.

On the other hand, if the angle between the vector from the object's gravity to the surface's center of gravity and the normal vector of the surface is 0, then the degree of freedom for rotation about the vector from the object's gravity to the surface's center of gravity is 0. remain. For this reason, we check whether the surfaces overlap based on the length and angle of the edges of the surfaces, and find the transformation that causes them to overlap by rotating around this axis.

(2) Check whether the transformation obtained in (1) is actually a symmetric transformation. That is, the transformation is applied to all the vertices of the object, and it is checked whether they all overlap with the vertices of the object before the transformation, and it is checked whether the transformation is symmetrical. It then returns a list of overlapping transformations.

(3) Delete duplicate conversions from the list obtained in (2).

In this way, it is possible to describe postures that have identity due to rotational symmetry.

Therefore, in determining the states, they can be treated as the same, and the grasping plan can be made appropriate.

In order to determine a posture for grasping a target object from a wide space in a grasping position vicinity spatial description creation work environment, a spatial description of the vicinity of the target object is created by grasping position vicinity spatial description creation means 5, and
The created spatial description is analyzed by the grasped position vicinity spatial description analysis means 6, and is described in the three-dimensional model 1 of the environment. First, a geodesic dome is created in which space is divided into pentagonal pyramids (spatial pyramids) with the grasped position as the apex. As shown in Figure 7, the number of spatial cones in this geodesic dome differs depending on the level, with level 0 having 20 planes, level 1 having 80 planes, level 2 having 320 planes, and level 3 having 1280 planes. There is.

Then, for the geodesic dome created in this way, a space cone (open space cone) whose distance from the apex of the space cone (grasping position) to interference with surrounding objects is a certain value or more is determined. That is, only geodesic domes created for one grasping position that do not interfere with surrounding objects are selected. When this process is performed for the state shown in FIG. 3(A), a space cone as shown in FIG. 8(A) is selected as the open space cone.

Next, on the surface of the geodesic dome, distance transformation is performed for each open space cone to examine the distribution of open space cones. That is, regarding the open space cone shown in FIG. 8(A) above, the distance 1 is set to the surrounding space cone adjacent to the non-open space cone (the space cone not shown in FIG. 8(A)). And this distance 1
Let the distance of the space cone adjacent to the space cone be 2. In this way, find the shortest distance for all open space cones.

By this, the shortest distance (distance conversion value) from all open space cones to the nearby non-open space cones is found.

This distance conversion value serves as an index regarding the size of the space as seen from the center of the geodesic dome. That is, the open space cone with the largest distance conversion value becomes the center of the direction in which the locally widest space exists. Therefore, by selecting an open space cone with a large distance conversion value, it is possible to determine the optimal grasping direction in one state. In this example, the eighth
The 16 open space cones shown in Figure (B) can be selected as those with the largest distance conversion values.

State Transition Determination Next, the state transition determination means 9 determines whether the target object can be transitioned from one state to another (big and place).

That is, when the state transition determining means 9 attempts to transition a target object in one state to another state, it determines whether there is a grasping posture that allows grasping the same place in the two states of the target object. If a grasping posture that allows grasping the same place exists, a transition is possible, and if it does not exist, a transition is impossible.

Here, this determination will be specifically explained.

First, grasping posture candidates that are likely to be grasped in each state are found. In other words, grasping neighborhood space description analysis means 6
The distance transformation value for the open space cone obtained in
Both of the grasping candidates calculated by the grasping posture candidate calculation means 2 are combined, and a distance conversion value corresponding to each grasping posture candidate is written. As a result, distance conversion values are written only for grasping postures with a high possibility of grasping.

At this time, in the case of an object with symmetry, since the sameness of grasping postures due to the symmetry is recognized as described above, it is possible to obtain grasping postures that take this into consideration.

Then, the distance conversion value for this grasping posture with a high possibility of grasping is described for both of the two target states, and the values are compared.

That is, between two states, grasping postures for which distance conversion values are both given have a high possibility of transition, and all of these are determined.

Here, the desired grasping posture is one that is larger than the converted value both at the time of pick and at the time of place.

Therefore, among the distance conversion values in the two states for each common graspable posture, the smallest distance conversion value is selected, and the selected values are arranged in ascending order.

Then, in this order, a manipulator operating range check and an interference check are performed, and those with no problems are selected.

This determines the grasping posture during the transition from one state to the next state. This determined grasping posture is one that can actually be grasped and is from the widest space.

In this way, the grasping posture can be determined for the transition from one state to another state.

However, if a direct transition from the initial state to the target state is not possible, it is necessary to change hands.

Therefore, we create a state transition sequence that takes into account changing hands.

Note that a known obstacle avoidance movement plan or the like is applied to the movement path of the robot.

State Transition Route Planning In this embodiment, the grasping plan and the planning of changing locations are performed separately in the process of creating a state transition series. That is, in the process of creating a state transition sequence, the state of the target object at the time of changing hands is represented only by a stable posture. Therefore, the state space of the target object consists only of the initial state, the target state, and a plurality of stable states.

At this time, since the above-mentioned rotational symmetry is also taken into account, the state space for objects with symmetry becomes smaller, and unnecessary searches can be omitted.

In the case of the L-shaped symmetrical object mentioned above, the state space for it has only eight states, as shown in FIG.

Then, the big-and-place operation is repeated in such a state space from the initial state to find the state transition sequence up to the target state, but in this case, instead of evaluating the state transitions for all combinations, , this search is performed using a method based on the A* algorithm of the state space search method.

Search Algorithm This search algorithm will be explained using graph symbols.

This graph is made up of a set of nodes.

It is assumed that the nodes are connected by oriented branches. If a branch is directed from node n to node m, then node m is said to be a continuation node of node n. Further, node n is called the parent node of node m.

When displaying the state space using a graph,
State descriptions are written at the nodes of the graph, and big-and-place operators are added to the edges.

The process of applying an operator to a node n and finding its continuation node is called expanding the node n.

In the search for the grip change problem using the state space as described above, there is a possibility that the target state and the stable posture state for grip change become continuous nodes by expanding the nodes.

Then, one continuation node with a high probability is created, and the possibility of creating subsequent continuation nodes is left to the later stages of the search. That is, first, the state transition determining means 9 attempts to transition to the goal node (target state).

After that, expansion is attempted one by one in the order of the number of stable postures, and the path with the lowest cost is selected. Note that for each contact point, an expansion count is defined as the number of times expansion to the continuation node is attempted in order to determine the next expansion state.

Here, in order to realize the A* algorithm, from the cost value g1 multiplied up to that point in any state, the predicted value h of the minimum cost related to the target state, and the cost evaluation value f-g If we define h and the cost of the actual shortest path from there to the goal state is h, then ho≧
It is necessary that h holds true.

Here, the number of pick-and-place operations is taken as the cost value. Therefore, h is 1 if there is a possibility of expanding to the goal node, and 2 if the node has already tried to expand to the goal node and failed. This allows the above h
Since the condition o≧h holds true, it is understood that the A* algorithm can always derive the shortest route.

Also, since all nodes have the possibility of being adjacent to the goal node, if there are multiple nodes where f is the same, there is a possibility that the goal node is a continuation node, that is, h is 1 Expand nodes preferentially. Furthermore, if there are several nodes with the same f and g, it is thought that there is a high possibility that they exist in a switching posture that has not been expanded before or on a path that is a solution, and priority is given to the node that has a larger expansion number. We are trying to expand it.

Next, specific steps will be explained.

(1) Put the start node in a list called 0PEN.

(2) If OPEN is empty, the process ends with failure. On the other hand, 0PE
If N is not empty, proceed to the next step.

(3) Take out the one with the lowest f from OP and EN.

If there are more than one, take out the one with h or 1 among them. If there are multiple h = 1, select the one with the largest number of expansions from among them. ).

The node selected here is called n.

(4) If n is the goal node, follow the pointer, find the solution route, and end with success. If it is not the goal node, proceed to the next step.

(5) Create a continuation node depending on the number of expansions and attempt expansion.

At this time, the number of expansions of the value continuation node is set to 0.

If the number of extensions of n is 0, an attempt is made to extend to the goal node; otherwise, an attempt is made to extend to a stable posture equal to the number of extensions. Increase the number of expansions of node n by one. When the number of expansions of node n becomes the number of stable postures + 1, it means that the expansion to all stable postures has been completed, so it is extracted from 0PEN and placed in a list called CLOSED. If expansion is not possible, return to (2) above.

(7) If the continuation node is not already on 0PEN or CLOSED, put it in 0PEN and attach a pointer from the continuation node back to n.

(8) If the continuation node is already on 0PEN, 5・PfI
Compare the Ili values f and leave the smaller node on 0PEN.

(9) If the continuation node is already on CLOSED, compare the evaluation value f, and if the new continuation node is smaller, change it to C
Take it from LOSED and put the new continuation node in 0PEN.

(10) Return to (2).

Here, the value of g at a certain node is determined from the number of nodes that have been traced so far by following the pointer. Also, the value of f is
If the number of extensions of that node is 0, it is 1 because there is a possibility of transitioning to the goal node in the next extension, and if the number of extensions is 1 or more, the goal node is aimed at after reaching a stable posture, so it is 2. Become.

In this way, a state transition sequence is determined by selecting the state to transition to while selecting the state with the smaller evaluation value f. Therefore, when reaching the goal, it is guaranteed that it is the shortest route.

After obtaining the state transition series, the holding change place is selected by the holding change place determining means 7.

The two-dimensional space within the robot's working range on the horizontal table plane is analyzed and candidates for the handhold changing location are selected in order from the widest space.

This will be explained based on FIG. 9. Figure 9 (A)
is an example of a two-dimensional space to be analyzed. Dots indicate the range in which the robot can take any grasping position on the table.

If a space in which no obstacle exists is selected, a space (dot) as shown in FIG. 9(B) is selected. Next, distance conversion is performed on the dots obtained in this manner from the surrounding areas, and the dot with the largest distance conversion value is selected. This is because a large distance conversion value means that there is the widest space around.

Therefore, two dots are selected as shown in FIG. 9(C), and these points are suitable locations for changing grips.

Next, a specific algorithm for calculating the change location will be explained.

(1) Divide the work space into small squares (here 1.5 cm) on the table and divide them into a square area where objects other than the object to be grasped exist (obstacle space) and a square area where there are no objects (free space) .

(2) Give an appropriate natural number N as the increment of the angle θ.

(3) For squares in free space on the table, perform distance conversion from the obstacle space, arrange the center positions (x, y) of these squares in descending order of distance conversion values, and create a strike L.

(4) If list L is empty, the process ends with failure. On the other hand, if list L is not empty, the first element is taken out from list L and assigned to X and Y. Also, θ is set as an initial value of O
Set it to .

(5) Update θ as θ−θ+360/N.

(6) Place the object to be grasped in a stable posture for changing hands at the location X, Y at an angle of θ, and check whether both the previously determined grasping posture for picking and the grasping posture for placing can be achieved. .

(7) If it is possible, end the process.

(8) If it is not possible, determine whether θ exceeds 360°. If θ is less than 360°, return to (5), update θ, and check again. On the other hand, if it exceeds 360 degrees, return to (4) and extract the next element from list L.

In this way, the location for changing hands is determined.

Example of a grip change operation plan A grip change operation plan for inserting the L-shaped object shown in FIG. 10 (A) into a hole as shown in FIG. 10 (F) was carried out using the method of the present invention. Here, candidates for grasping postures of the target L-shaped object are shown in FIG. There are a total of 138 grasp posture candidates.

According to the plan created by the device of the invention, the 10th
Transition occurs to the states shown in Figures (A) to 10 (F).
Since the object in Figure (A) is located near an obstacle, the posture that can be grasped is limited. The grasping posture is limited to one by the grasping posture determination method described above. Next, the state moves to the state shown in FIG. 10(B) according to the state transition path. As mentioned above, the selection of this handholding location is set to the location least affected by obstacles. Then, the target object is changed according to the path determined by the state transition path (Fig. 10 (C)).
), as shown in Figure 10 (D), the stable posture in Figure 4 (
B) will be changed. From this state, Figure 10 (E
) and transitions to the target state shown in FIG. 10(F).

As described above, in this example, the work cannot be performed without changing the grip twice, and the states shown in FIGS. 4(G) and (B) are selected as stable states for changing the grip.

[Effects of the Invention] As explained above, according to the robot work planning device according to the present invention, the distance in each open space cone obtained by the distance conversion means and the grasping candidate obtained by the grasping posture candidate calculation means are calculated. Since the grasping posture is determined based on both, efficient grasping posture determination can be achieved.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of a robot work planning device according to the present invention; FIG. 2 is an overall flowchart for explaining the operation of the embodiment; FIG. 3 is an explanatory diagram showing grasping posture candidates; Figure 4 is an explanatory diagram showing a stable posture, Figure 5 is an explanatory diagram showing rotational symmetry of an object, Figure 6 is an explanatory diagram when grasping an object with rotational symmetry, and Figure 7 is an explanatory diagram showing a geodesic dome. FIG. 8 is an explanatory diagram showing a spatial description in the vicinity of the grasping position, FIG. 9 is an explanatory diagram for explaining selection of a holding position, and FIG. 10 is an explanatory diagram showing an example of a work plan. 1...Three-dimensional model of the environment 2...Grasp posture candidate calculation means 3...Stability posture calculation means 4...Rotational symmetry calculation means 5...Grasp position vicinity spatial description creation means 6...
Grasping position vicinity spatial description analysis means 7 ... Handover location determination means 8 ... Work plan creation means 9 ... State transition determination means

Claims (1)

[Scope of Claim] A robot work planning device that plans a grasping operation of an object by a robot from a three-dimensional environment model including an initial state of the object and states of surrounding objects, comprising: , a grasping posture candidate calculation means for calculating a grasping posture candidate that is a posture capable of grasping the object at a predetermined grasping position of the object based on the shape of the object grasping part of the robot; an open space cone selection means for dividing the space into space cones of a predetermined size with the grasping position as the apex, and selecting an open space cone that does not interfere with surrounding objects from among the space cones; a distance conversion means for determining the distance from surrounding objects for each of the cones; a distance in each open space cone obtained by the distance conversion means;
A robot work planning device comprising: grasping posture determining means for determining a grasping posture based on both grasping posture candidates obtained by grasping posture candidate calculation means.